APPLIED AND ENVIRONMENTAL MICROBIOLOGY, June 2011, p. 4155–4162 Vol. 77, No. 120099-2240/11/$12.00 doi:10.1128/AEM.02847-10Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Biocontrol of Ralstonia solanacearum by Treatment withLytic Bacteriophages�†

Akiko Fujiwara, Mariko Fujisawa, Ryosuke Hamasaki, Takeru Kawasaki,Makoto Fujie, and Takashi Yamada*

Department of Molecular Biotechnology, Graduate School of Advanced Sciences of Matter, Hiroshima University,Higashi-Hiroshima 739-8530, Japan

Received 6 December 2010/Accepted 9 April 2011

Ralstonia solanacearum is a Gram-negative bacterium and the causative agent of bacterial wilt in manyimportant crops. We treated R. solanacearum with three lytic phages: �RSA1, �RSB1, and �RSL1. Infectionwith �RSA1 and �RSB1, either alone or in combination with the other phages, resulted in a rapid decreasein the host bacterial cell density. Cells that were resistant to infection by these phages became evidentapproximately 30 h after phage addition to the culture. On the other hand, cells infected solely with �RSL1in a batch culture were maintained at a lower cell density (1/3 of control) over a long period. Pretreatment oftomato seedlings with �RSL1 drastically limited penetration, growth, and movement of root-inoculatedbacterial cells. All �RSL1-treated tomato plants showed no symptoms of wilting during the experimentalperiod, whereas all untreated plants had wilted by 18 days postinfection. �RSL1 was shown to be relativelystable in soil, especially at higher temperatures (37 to 50°C). Active �RSL1 particles were recovered from theroots of treated plants and from soil 4 months postinfection. Based on these observations, we propose analternative biocontrol method using a unique phage, such as �RSL1, instead of a phage cocktail with highlyvirulent phages. Using this method, �RSL1 killed some but not all bacterial cells. The coexistence of bacterialcells and the phage resulted in effective prevention of wilting.

Bacterial wilt is an important crop disease, caused by thesoilborne Gram-negative bacterium Ralstonia solanacearum.This bacterium has an unusually wide host range, infectingmore than 200 species belonging to more than 50 botanicalfamilies, including economically important crops (9, 10). R.solanacearum strains represent a heterogeneous group subdi-vided into five races based on host range, five biovars based onphysiological and biochemical characteristics (8), and four phy-lotypes roughly corresponding to geographic origins. PhylotypeI includes strains originating primarily from Asia, phylotype IIfrom America, phylotype III from Africa and surrounding is-lands in the Indian Ocean, and phylotype IV from Indonesia(4). In the field, R. solanacearum is easily disseminated via soil,contaminated irrigation water, surface water, farm equipment,and infected biological material (14). Bacterial cells can sur-vive for many years in association with alternate hosts, and soilfumigation with methyl bromide, vapam, or chloropicrin is oflimited efficacy. Because methyl bromide depletes the strato-spheric ozone layer, the production and use of this gas wasphased out in 2005 under the Montreal Protocol and the CleanAir Act. Due to the limited effectiveness of the current inte-grated management strategies, bacterial wilt continues to be aneconomically serious problem for field-grown crops in manytropical, subtropical, and warmer areas of the world (9, 10).

Like other methods of biological control, one advantage ofphage therapy (also called phage biocontrol) is the reductionin the use of chemical agents against pathogens. This avoidsproblems associated with environmental pollution, ecosystemdisruption and residual chemicals on crops. Phage therapy inagricultural settings was extensively explored 40 to 50 years agoas a means of controlling plant pathogens (3, 18). Two majorproblems arose in those trials: (i) extracellular polysaccharidesproduced by pathogenic bacteria prevented phage adsorption,and (ii) there were various degrees of susceptibility amongbacterial strains (8). Nevertheless, over recent decades, the useof phage therapy to control the growth of plant-based bacterialpathogens has been explored with increased enthusiasm. Tocontrol R. solanacearum, two bacteriophages have alreadybeen isolated, and their physical and physiological propertieshave been characterized: phage P4282 (19, 21) and phagePK101 (22). Both of these phages demonstrate very narrowhost ranges and infect only a few strains of R. solanacearum.Phage P4282, which infects R. solanacearum strain M4S, wasused to control bacterial wilt in tobacco plants under labora-tory conditions, and possible phage-mediated protection wasobserved (21). However, for the practical use of phages asbiocontrol agents against bacterial wilt, it is believed that mul-tiple phages with wide host ranges and strong lytic activity arerequired (22). Recently, Yamada et al. (23) isolated and char-acterized several different kinds of phage that specifically in-fected R. solanacearum strains belonging to different racesand/or biovars. Phage �RSA1 is a P2-like head-tail virus (Myo-viridae) with a very wide host range; all 15 strains tested fromrace 1, 3, or 4 and biovar 1, N2, 3, or 4 were susceptible to thisphage (6). Phage �RSL1 is another myovirus containing �231kb in its genome; this phage was able to lyse 10 of 15 tested

* Corresponding author. Mailing address: Department of MolecularBiotechnology, Graduate School of Advanced Sciences of Matter, Hir-oshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8530,Japan. Phone and fax: 81-82-424-7752. E-mail: tayamad@hiroshima-u.ac.jp.

† Supplemental material for this article may be found at http://aem.asm.org/.

� Published ahead of print on 15 April 2011.

4155

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

strains (24). The most recently isolated phage, �RSB1, dis-played a T7-like morphology (Podoviridae) and also had thewidest host range, with 14 of 15 strains from race 1, 3, or 4being susceptible (17). �RSB1 lyses host cells and forms verylarge clear plaques that are 10 to 15 mm in diameter on assayplates.

The three phages—�RSA1, �RSB1, and �RSL1—appearto be useful in the eradication of the bacterial wilt pathogen.To increase the antibacterial efficacy of these phages in bio-control, a thorough understanding of phage ecology and com-plex phage-host interactions in various environments is neces-sary. Genomic information on these phages and their hostbacteria will be useful for understanding the phage character-istics and the history and molecular mechanisms involved inthe phage-bacterium interactions.

MATERIALS AND METHODS

Bacterial strains and phages. Strains of R. solanacearum were obtained fromthe culture collections as listed in Table 1 . The avirulent strain M4S was used forroutine purposes (21). For plant inoculation, a few virulent strains, such as Ps29and MAFF 730138, were used. Strain MAFF 106611 was used mainly for inplanta detection of bacterial cells. The bacterial cells were cultured in CPGmedium containing 0.1% Casamino Acids, 1% peptone, and 0.5% glucose (11)at 28°C with shaking at 200 to 300 rpm. Bacteriophages �RSA1, �RSB1, and�RSL1 (Table 2) were routinely propagated using strain M4S as the host.Bacterial cells in the stationary phase (16 to 24 h postinoculation) grown in CPGmedium were diluted 100-fold with 100 ml of fresh CPG medium in a 500-mlflask. To collect sufficient phage particles, a total of 1 liter of bacterial culture wasgrown. When the optical density at 600 nm (OD600) of the cultures reached 0.1to 0.3, the phage was added at a dose of 0.5 PFU/host cell (0.5 � 108 PFU/ml)for �RSA1 and �RSB1 or 5.0 PFU/cell (5 � 108 PFU/ml) for �RSL1. Afterfurther growth for 6 to 24 h, the cells were removed by centrifugation in a HitachiHimac CR21E centrifuge equipped with a R12A2 rotor (Hitachi Koki Co., Ltd.,Tokyo, Japan) at 8,000 � g for 20 min at 4°C. To increase �RSA1 recovery,EGTA (final concentration, 1 mM) was added to the �RSA1-infected culture at6 to 9 h postinfection (p.i.). The supernatant was passed through a 0.45-�m-pore-size membrane filter, and phage particles were precipitated by centrifuga-tion in a Hitachi CP100� centrifuge with a P28S rotor at 40,000 � g for 1 h at 4°Cand then suspended in SM buffer (50 mM Tris-HCl [pH 7.5], 100 mM NaCl, 10mM MgSO4, 0.01% gelatin). Purified phages were stored at 4°C until needed.

DNA manipulations. Standard molecular biological techniques for DNA iso-lation, digestion with restriction enzymes and other nucleases, and constructionof recombinant DNAs were as described by Sambrook and Russell (20). PhageDNA was isolated from the purified phage particles by phenol extraction (2, 23).

In planta detection of R. solanacearum cells. Seeds of the tomato (Lycopersiconesculentum) cultivar Oogata-Fukuju were obtained from Takii Co., Ltd. (Kyoto,Japan). For aseptic cultures, seeds were surface sterilized with sodium hypochlo-rite and cultured in a square dish (sterile square Schale no. 2; Eiken ChemicalCo., Ltd., Tokyo, Japan) containing solid medium (0.15% Hyponex powder;Hyponex Japan Corp., Ltd., Osaka, Japan), 0.5% sucrose, and 1.5% agar ad-justed to pH 5.8. Plants were grown in an incubator (Sanyo growth cabinet;Sanyo, Osaka, Japan) at 28°C under a 16-h light (300 �mol photons/s/m2) and8-h dark cycle. During the culture period, the dishes in the chamber were tiltedto a 45° angle to encourage roots to grow on the surface of the medium. Thismade it easy to access the roots and monitor the inoculated bacterial cells under

a stereomicroscope. To inoculate plants, bacterial cells (i.e., strain MAFF 106611bearing pRSS12, a green fluorescent protein [GFP]-expressing plasmid with noeffects on host virulence [5, 16]) were cultured in CPG medium and suspendedin sterile distilled water at a density of 107 to 108 cells/ml. Tomato seedlingsgrown in culture dishes were cut at the tip of the taproot, 10 mm from the apex,with a razor blade and then pretreated with phages; 1.0 �l of phage preparation(109 PFU/ml) was added to the cut. For a mock control, 1.0 �l of distilled waterwas added. After 12 to 24 h from the phage treatment, 1.0 �l of bacterialsuspension (107 to 108 cells/ml) was applied to the section. After inoculation, theplants in the dishes were cultured in the incubator until observation. Bacterialcells in the plants were observed by using an MZ16F fluorescence stereomicro-scope (Leica Microsystems, Heidelberg, Germany) equipped with GFP2 andGFP3 filters and/or an Olympus BH2 fluorescence microscope (Olympus, Tokyo,Japan). Microscopic images were recorded with a charge-coupled device camera(VB-6010; Keyence, Osaka, Japan). For prolonged observation of plants, bacte-rium-treated plants were transferred from culture dishes to pots containing amixture of peat moss and expanded vermiculite. Plants were grown under naturalconditions.

Detection of �RSL1 remaining in plants and soil. Two surviving plant cul-tures, 4 months after treatment, were subjected to assays to detect any remainingbacteriophages. Two parts of the stem (2.5 g at 10 cm above the soil surface and0.7 to 0.8 g at 1 to 5 cm above the soil surface) were excised from the plant. Forthe root, 2 to 3 g (12 to 13 cm) was removed after washing the sample in runningtap water. After the addition of 2 ml of CPG medium, the plant material wasground using a mortar and pestle at room temperature. After centrifugation at4,200 � g for 5 min at 4°C, the supernatant was filtered through a 0.45-�m-pore-size membrane filter (Steradisc; Krabo Co., Okayama, Japan). The filtrate(100-�l aliquot) was subjected to plaque assays with strain Ps29 as the host onCPG plates containing 0.45% agar. A 2-g mass of pot soil was suspended in 5 mlof SM buffer, and then, after mixing and filtration through a membrane filter asdescribed above, a 100-�l aliquot of the suspension was subjected to plaqueassays.

Treatment of tomato plants in soil with phages and inoculation with R.solanacearum cells. Tomato seeds (cv. Oogata-Fukuju) were planted in Jiffy 7peat pellets (42 mm in diameter; Sakata Seed Co., Ltd., Kanagawa, Japan), whichwere soaked with a phage solution (1.3 � 1010 PFU/pot). For controls, pelletswere soaked with tap water. After cultivation for 1 month, the plants (20 to 23 cmin height) were treated again with the phage solution (1.3 � 1010 PFU/pot); thepellet was soaked with the phage solution. Two days later, the plants were cut atthe root tips with scissors and dipped in a bacterial suspension containing strainMAFF 106611 cells at 108 cells/ml for 30 s. The bacterium-treated plants weretransferred to pots (9 cm in diameter) containing peat moss and expandedvermiculite and grown in an incubator (Sanyo) at 28°C under a 16-h light and 8-hdark cycle. Symptoms of wilting were graded from 0 to 5 as follows: 0, nosymptoms; 1, only one petiole is wilting; 2, two to three petioles were wilting; 3,all but two to three petioles were wilting; 4, all petioles were wilting; and 5, theplant died.

Effect of temperature on the stability of �RSL1. Phage preparations of�RSA1 (109 PFU/ml), �RSB1 (109 PFU/ml), and �RSL1 (109 PFU/ml) in SMbuffer (0.5 ml) were incubated in sealed tubes at 4, 11, 28, 37, and 50°C forvarious periods before the plaque assay with strain Ps29 as the host. A 10-mlvolume of each phage solution was added to 100 g of autoclaved field soil. Thephage-soil mixture was divided into five equal amounts and separately filled into15-ml tubes. After sealing, the tubes were incubated at 4, 11, 28, 37, and 50°C forvarious periods before the plaque assay was carried out. These experiments wereperformed twice.

RESULTS

Treatment of R. solanacearum cells with three differentphages. R. solanacearum M4S cells were treated with three

TABLE 2. Bacteriophages used in this study

Bacteriophage Family Type Genomesize (bp) Reference

�RSA1 Myoviridae P2-like 38,760 6�RSB1 Podoviridae T7-like 43,077 17�RSL1 Myoviridae Jumbo phage 231,255 24

TABLE 1. Bacterial strains used in this study

R. solanacearumstrain

DescriptionSourcea

Race Biovar Phylotype

M4S 1 3 I LTRCPs29 1 3 I LTRCMAFF 106611 1 4 I NIASMAFF 730138 1 3 I NIAS

a LTRC, Leaf Tobacco Research Center, Japan Tobacco, Inc. (23); NIAS,National Institute of Agrobiological Sciences, Japan (23).

4156 FUJIWARA ET AL. APPL. ENVIRON. MICROBIOL.

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

lytic phages—�RSA1, �RSB1, and �RSL1—alone or in com-bination. The optimal phage doses per bacterial cell to give thehighest phage yield were determined to be 0.5, 0.5, and 5.0 for�RSA1, �RSB1, and �RSL1, respectively. Figure 1 shows theeffect of phage infection on bacterial growth. �RSA1 and�RSB1 infection, irrespective of whether they were used solelyor mixed with other phages, were able to readily lyse growingcultures. However, at approximately 30 h p.i., resistant cellsstarted to grow and reached the stationary state at 70 h p.i. Theresults were the same at different phage doses (PFU/bacterialcell) for each of the phages. Recovering cells demonstrated ageneral resistance to both �RSA1 and �RSB1, irrespective ofthe initial phage species they were infected with (Table 3). Thegrowth recovery pattern at 30 h p.i. was not affected by theaddition of other phage species at 40 h p.i. (data not shown). Incontrast, cells infected solely with �RSL1 grew slowly until 40to 50 h p.i., instead of quickly lysing, and were maintained ata low and steady level until 60 h p.i. (Fig. 1). This steady-state growth pattern may indicate an equilibrium betweencell growth and lysis by the phage or resistant bacterial cellsthat are less robust in their growth (1). Infection of �RSL1in combination with �RSA1 and/or �RSB1, however, re-sulted in cell lysis and recovering growth patterns similar tothat seen in bacterial cells infected solely with �RSA1, �RSB1,or a mixture. Our previous results have shown that these threephages are virulent and form clear plaques on culture plates(23). Both �RSA1 and �RSB1 lyse the host cells at 2 to 3 h p.i.(6, 17), whereas �RSL1 takes longer for lysis to occur (3 to 4 h)(24). Therefore, one possible explanation for the observationin the mixed infection with �RSL1 is that �RSA1 and/or�RSB1 predominantly infected and lysed the host cells, andover time the titers of these phages were greatly increasedcompared to that of �RSL1, resulting in the lysis-recoverypattern of cell growth (Fig. 1). We confirmed this throughplaque assays with the culture fluid after 100 h p.i. In the caseof the �RSL1 and �RSA1 mixed infection, 74% of plaques(�108) formed on the plates were due to �RSA1, easily dis-tinguishable by plaque morphology. For the culture coinfectedwith �RSL1 and �RSB1, 92% of plaques (2 � 108) were dueto �RSB1. �RSB1 was also predominant (�90% of 9 � 108

plaques) when combined with �RSA1. Phage-resistant cellsfrom cultures infected with �RSA1 and �RSB1 were sus-ceptible to �RSL1, as shown in Table 3. Therefore, �RSL1presumably has a different host recognition way from theother phages. These results obtained with strain M4S as ahost were reproducible with other R. solanacearum strainssuch as Ps29 and MAFF 730138.

From these results, we concluded that for control R. so-lanacearum cells, a cocktail of �RSA1, �RSB1, and �RSL1may not be suitable due to the presence of phage-resistantcells. Instead, �RSL1 infection somehow maintains the cellpopulation at lower levels and limits the growth of resistantcells recovering from �RSA1 and/or �RSB1 infection. There-fore, we attempted to use �RSL1 as an agent to control thegrowth of R. solanacearum cells and bacterial wilt disease.

In planta inhibition of R. solanacearum growth and move-ment by treatment with �RSL1. The effect of �RSL1 infectionto stably limit the growth of R. solanacearum cells in vitro leadsus to examine bacterial growth and movement in planta aftertreatment with �RSL1. We monitored real-time bacterial dy-

namics in inoculated tomato plants grown on solid agar me-dium using a previously described method (5, 16). Seven-day-old seedlings of the tomato cultivar Oogata-Fukuju were firsttreated with �RSL1 (1.0 �l containing 106 PFU) at a cut madein the tip of the taproot, 10 mm from the apex. At various timesafter treatment, bacterial cells were inoculated at the cut andobserved (Fig. 2A). Every experiment was performed in trip-licate with 10 individual plants. Figure 2B shows the timecourse of bacterial growth and movement in a tomato taprootwithout �RSL1 treatment (control). After bacterial inocula-tion, GFP fluorescence intensity increased with time, andGFP-labeled bacterial cells moved upward through xylem ves-sels until 48 h p.i. At 72 h p.i., GFP fluorescence was apparentoutside the taproot, suggesting cell movement, and growthoccurred outside the taproot. Slimy colonies of cells coveredthe entire taproot by 96 h p.i. At this stage, the hypocotyls andyoung leaves were wilting (Fig. 2B).

Remarkably different phenomena were observed when to-mato plants were treated with �RSL1 preceding bacterial in-oculation. As shown in Fig. 2C, up until 96 h p.i. no bacterialgrowth and movement into plant bodies were apparent; faintGFP fluorescence was retained at the inoculation point. Dur-ing this period, lateral shoots frequently formed as in healthyplants; this was never observed in control plants that werebacterially challenged but not phage treated. At 120 h p.i., GFPfluorescence intensity increased slightly, and GFP-labeled cellswere observed to be moving upward along the taproot. How-ever, the growth and movement was quite limited; at 192 h p.i.,the GFP fluorescence pattern remained unchanged. The GFPfluorescence was unclear again at 216 h p.i. The cotyledons,leaves, and the meristem looked healthy with no symptoms ofwilting (Fig. 2C). When tomato seedlings were treated in thesame way with �RSA1 or �RSB1, bacterial penetration andgrowth in xylem vessels were obvious at 48 h p.i. (data notshown). These results indicated that bacterial growth andmovement were drastically limited in tomato plants by pre-treatment with �RSL1. Thus, pretreatment of tomato seed-lings with �RSL1 may limit infection by R. solanacearum orprevent bacterial wilt disease.

Prolonged cultivation in soil of tomato plants treated with�RSL1 after inoculation with R. solanacearum. Six tomatoplants were treated with �RSL1 after inoculation with R. so-lanacearum (Fig. 2), transferred from culture dishes to pots,and then cultivated in a natural environment for prolongedobservation. At 49 days posttransfer (1.5 months p.i.), fourplants still grew well without any symptoms of wilting. Twoplants died soon after transfer (2 to 3 days); we considered thatthis was caused by physical/physiological damage such as watershock or root damage and not by bacterial infection becausefew bacterial cells were detected in plant organs.

Persistence and stability of �RSL1 particles in plants andsoil. The stems, leaves, and roots of plants and culture soilfrom two pot cultures described above were subjected to phagetiter assays 4 months later. One plant, 52 cm in height, wastreated with �RSL1 only (control plant 1), and the other plant,62 cm in height, was treated with �RSL1 followed by thepathogen (plant 2), and harvested 4 months p.i. The datashown in Table 4 indicated that no �RSL1 phage was detectedfrom the stems or leaves of either plant. However, a consider-ably large number of phages were retained in the roots of both

VOL. 77, 2011 PHAGE BIOCONTROL OF R. SOLANACEARUM 4157

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

plants. The plants inoculated with bacterial cells gave titers 10times higher than those without inoculation. Similar resultswere obtained in the in vitro experiments shown in Fig. 2). Thesoils from both cultures also yielded phage plaques. Plaques

appearing on the assay plates always displayed homologousmorphology, resembling �RSL1 plaques (23). The genomicDNA isolated from random plaques coincided with �RSL1DNA by restriction digestion patterns (data not shown). These

FIG. 1. Time course of bacterial growth after infection with bacteriophages. The first 20 h region is enlarged in panel B (the symbols are thesame as shown in panel A). Cells of R. solanacearum strain M4S (OD600 � 0.3 corresponding to �108 cells/ml; vertical arrow) were infected solelyor by mixing with three phages: �RSA1, dose � 0.5 � 108 PFU/ml; �RSB1, dose � 0.5 � 108 PFU/ml; and �RSL1, dose � 5.0 � 108 PFU/ml.At about 30 h p.i., resistant cells started to grow when treated with �RSA1 and/or �RSB1, either alone or as part of a phage mixture. Cells solelyinfected with �RSL1 were kept at a low cell density. Similar results were obtained with different R. solanacearum strains (data not shown).

4158 FUJIWARA ET AL. APPL. ENVIRON. MICROBIOL.

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

results indicated that �RSL1 phages were stably retained inplants roots, as well as in soils, with a lower concentration.

Prevention of bacterial wilt by treatment with �RSL1. Theeffect of �RSL1 infection to stably limit the growth of R.solanacearum cells in vitro and in planta led us to examinewhether �RSL1 treatment of tomato plants under soil condi-tions prevents wilting by inoculated R. solanacearum cells.One-month-old tomato plants (20 to 23 cm in height) pre-treated with �RSL1 were inoculated with R. solanacearumcells as described in Materials and Methods. Wilting symptomswere recorded every 2 days. The results shown in Fig. 3 indi-cated an efficient prevention of wilting by treatment with�RSL1. Plants without �RSL1 started to show wilting symp-toms 4 days p.i. and all 11 plants showed wilting symptoms 18days p.i. (Fig. 3A and C). In contrast, all 11 �RSL1-treatedplants showed no symptoms of wilting during the experi-

mental period (Fig. 3B and C). The �RSL1 phages werestably retained in plant roots as well as in soils at this stage,as shown in Table 4. Three random tomato cultures treatedwith �RSL1 were subjected to phage titer assay as describedabove. Because all three cultures showed almost the samevalues, only one example (plant 3) is included in Table 4. Theseresults suggest the effectiveness of �RSL1 in the practical

TABLE 3. Phage resistance of R. solanacearum cells in a culturetreated with �RSA1 or �RSB1

Single colonyaPhage resistanceb

�RSA1 �RSB1 �RSL1

Ar-1 � � –Ar-2 � � –Ar-3 � � –Br-1 � � –Br-2 � � –Br-3 � � –

a Data for the single colonies isolated from �RSA1- or �RSB1-treated cul-tures at 100 h p.i. shown in Fig. 1 are presented. Ar-1, Ar-2, and Ar-3 and Br-1,Br-2, and Br-3 were obtained from cultures treated with �RSA1 and �RSB1,respectively.

b Phage resistance is indicated as resistant (�) or sensitive (–).

FIG. 2. In planta inhibition of R. solanacearum growth and movement by treatment with �RSL1. (A) Tomato seedlings grown in culture dishes(7 days old) were cut at the tip of the taproot (arrow) and treated with 1 �l of �RSL1 (106 PFU). After 12 to 24 h, bacterial cells (�105 cells) ofstrain MAFF 106611 harboring pRSS12 were inoculated at the cut. (B) Without phage treatment (control), bacterial penetration into the taproot,successive upward movement, and growth in the tissues was evident at 24 h p.i. (left panel, bright-field image; right panel, dark-field image). GFPfluorescence was apparent outside the taproot at 72 h p.i. The hypocotyl and young leaves were wilted at 96 h p.i. No lateral shoot formation wasobserved throughout the period. (C) With phage treatment, no bacterial growth and movement in plant bodies was apparent until 96 h p.i. FaintGFP fluorescence was retained at the inoculation point. During this period, lateral shoots frequently formed; even though at 120 h p.i. someGFP-labeled cells were visible, their growth and movement were quite limited. The cotyledon, leaves, and the meristem appeared healthy with nosymptoms of wilting at 288 h p.i. The 30 different plants tested displayed similar results. Scale bar, 0.5 mm.

TABLE 4. Persistence and stability of �RSL1 particles in plantsand soil

Plant or soila Wt (g) Mean phage titerb

(PFU/g) SD

Plant 1Stems and leaves (10 cm) 2.50 NDStems (1 to 5 cm parts

above ground)0.75 ND

Roots (13 cm) 2.50 1.8 � 103 2.0 � 102

Soil (rhizosphere) 2.00 2.7 � 101 8.5 � 100

Plant 2Stems and leaves (10 cm) 2.50 NDStems (1 to 5 cm parts

above ground)0.80 ND

Roots (13 cm) 3.10 1.9 � 104 2.5 � 103

Soil (rhizosphere) 2.00 5.5 � 102 1.5 � 102

Plant 3Roots 0.51 5.7 � 105 3.0 � 104

Soil (rhizosphere) 2.00 7.1 � 105 1.8 � 105

a Plant 1, a control plant treated with �RSL1 only (in vitro treated and trans-ferred to soil); plant 2, a plant treated with �RSL1, followed by the pathogen (invitro treated and transferred to soil); plant 3, a �RSL1-treated plant with nowilting symptom 18 days after the pathogen challenge (in soil experiments).

b Assays were repeated three times. ND, not detected.

VOL. 77, 2011 PHAGE BIOCONTROL OF R. SOLANACEARUM 4159

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

application as a biocontrol agent against bacterial wilt. Plantscan be treated with �RSL1 during the early stages of growth.When tomato plants were pretreated with a phage mixturecontaining two phages (�RSA1 and �RSB1) or three phages(�RSA1, �RSB1, and �RSL1) and challenged by R. so-lanacearum cells in the same way, as described above, wiltingsymptoms appeared by 4 days p.i., and all of them wilted after14 days p.i. (see Fig. S1 in the supplemental material). Thewilting patterns were generally similar to those observed incontrol plants without phage treatment. These results wereconsistent with the growth of resistant cells observed in the invitro and in planta experiments in which a phage mixture wasused.

Effects of temperature on the stability of �RSL1 in soil. Wefurther studied the effects of temperature on the stability of�RSL1 compared to �RSA1 and �RSB1. Phage preparations

were kept at different temperatures in the range 4 to 50°C, inthe presence or absence of soil. As shown in Fig. 4A, in theabsence of soil, all three phages were essentially stable below28°C, whereas �RSL1 exhibited greater stability compared toother phages at higher temperatures (37 and 50°C). After a15-day incubation, ca. 10% �RSL1 survived at 50°C. Nophages were detected after a 3-day incubation with �RSA1 andafter a 9-day incubation with �RSB1. Similar stability patternswere observed under different soil temperatures. At lowtemperatures, the titers of some phages, especially �RSA1,were decreased probably due to nonspecific adsorption tosoil particles. �RSL1 also exhibited the highest stability insoil (Fig. 4B).

DISCUSSION

Alternative phage biocontrol using �RSL1. For phage bio-control, only virulent phages are used, thereby avoiding theproblem of lysogeny. Phage cocktails are recommended toprevent the problem of resistance. These cocktails ideally con-tain several phages with different host specificities, replicationmechanisms, and/or infection cycles (7, 15). In the presentstudy, when host R. solanacearum cells were quickly lysed bytreatment with �RSA1 or �RSB1, resistant cells presumablypreexisting in the population at a very low frequency wereincreased at 30 h p.i. Given that the majority of the cell pop-ulation contains susceptible cells, the elimination of these cellsmay allow minor cells to predominate to the next generation.Since such recovering cells were somehow resistant to both�RSA1 and �RSB1 (Table 3), treatment with a mixture ofthese phages resulted in the same death and recovery patternsas bacterial cells treated with a single phage (Fig. 1). Theresistance mechanisms used by these cells are currently un-known, and the host specificity was different between �RSA1(a P2-like myovirus) and �RSB1 (a T7-like podovirus) (23). Inour preliminary observations, several phage-resistant mutants,induced by transposon mutagenesis, displayed differences inresistance to �RSA1 and �RSB1. Therefore, cells recover-ing from the phage treatment may include resistant cellscaused by different mechanisms and with different charac-teristics. A cocktail containing three phages—�RSA1,�RSB1, and �RSL1—also failed to stably prevent bacterialgrowth (Fig. 1). Although �RSL1 could lyse some recoveringcells after �RSA1 and/or �RSB1 treatment (Table 3), in mixedphage treatments, cells may have been quickly lysed by �RSA1or �RSB1, thereby interrupting the replication of �RSL1 andreducing the �RSL1 titer. �RSL1 may not be able to effect thelysis of cultures if its density is too low. Compared to �RSA1and �RSB1, RSL1 takes longer for lysis to occur (3 to 4 h) witha latent period of 2.5 h (24). This infection cycle is ratherlonger than the doubling time of host cells (3 h) under routineculture conditions. Therefore, a higher dose may be requiredto efficiently infect and lyse the host cells. This was supportedby the observation that little �RSL1 was detected in the cul-ture treated with three phages at 100 h p.i.

Consequently, the strategy of using a phage cocktail in thebiocontrol of R. solanacearum cells did not work in the presentstudy. Instead, treatment with �RSL1 alone did not rapidly killcells but kept the cell density at a low level (Fig. 1). In plantamonitoring of bacterial cells showed that treatment with

FIG. 3. Prevention of bacterial wilt by treatment with �RSL1. One-month-old tomato plants (20 to 23 cm in height) pretreated with tapwater (A, control) or �RSL1 (B) were inoculated with R. solanacearumcells as described in Materials and Methods. Wilting symptoms weregraded from 0 to 5 as follows: 0, no symptoms; 1, only one petiole waswilting; 2, two to three petioles were wilting; 3, all but two to threepetioles were wilting; 4, all petioles were wilting; and 5, the plant died.(C) Tomato plants observed at 18 days p.i.

4160 FUJIWARA ET AL. APPL. ENVIRON. MICROBIOL.

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

FIG. 4. Effect of temperature on the stability of �RSL1 compared to �RSA1 and �RSB1. Phage preparations were kept at a range oftemperatures (4 to 50°C) for various periods without (A) or with (B) soil. �RSL1 displayed significant stability at higher temperatures (37 and50°C). Error bars indicate the standard error (n � 3).

4161

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

�RSL1 resulted in a blockage of the growth and movement ofbacterial cells in the tomato root. Treated plants survived foras long as 4 months. �RSA1 or �RSB1 were not able to inducesimilar plant-protecting effects. Based on these observations,we propose an alternative phage biocontrol method, using aunique phage such as �RSL1, instead of a phage cocktailcontaining highly lytic phages. With this method, bacterial cellsare not killed altogether but a sustainable state of phage-bacterium coexistence is maintained, like a carrier state orpseudolysogeny (1). To apply this method practically, variantsof �RSL1-type phages would be required that have differenthost ranges covering most strains of R. solanacearum.

Coexistence of R. solanacearum cells and �RSL1. �RSL1 isa unique phage with a very large genome of 231 kb containing343 open reading frames. A lysogenic cycle or episomal repli-cation of �RSL1 has not been elucidated; genomic Southernblot analysis of many field-isolated strains has not identifiedany significant hybridizing signals with �RSL1 DNA as a probe(23; data not shown). After infection with �RSL1, cell densitywas maintained stably at low levels, as shown in Fig. 1, sug-gesting an equilibrium between cell growth and lysis. The es-tablishment of equilibrium may be explained by a long latentperiod and small burst size of a phage. Our previous studyrevealed that �RSL1 takes longer for lysis to occur (3 to 4 h)with a latent period of 2.5 h and a burst size of 80 to 90 PFUper infected cell (24). Other possible explanations for the cell-phage balance may come from some specific regulation ofphage infectivity or a low frequency of resistant cells and theproduction of growth-inhibitory factors due to phage infection.We are interested in the expression of many unique genes,including those for lysis genes encoded in the �RSL1 genomicDNA (24). Further characterization of the expression andfunction of these genes may reveal the mechanism for estab-lishment of equilibrium. It is also interesting that in somecases, lysogenic filamentous phages were observed to be in-duced in the remaining host cells after infection with �RSL1,which may also contribute to the equilibrium.

Stability of �RSL1 in soil and its practical application.Phages are utilized for controlling plant pathogens either inthe rhizosphere or phyllosphere. The direct application ofphages to the phyllosphere is subject to serious phage stabilityproblems (15). Field and laboratory studies have demonstratedthat phages are inactivated rapidly by exposure to sunlight,high temperatures, extremes in pH, oxidative conditions, andflowing water (12, 13). In the case of bacterial wilt, phages forbiocontrol can be applied to the rhizosphere. Sunlight, themost destructive environmental factor, and oxidative inactiva-tion are not as relevant in this case. �RSL1 was shown to berelatively stable in soil, especially at higher temperatures (Fig.4). In fact, active �RSL1 particles were consistently recoveredfrom the roots of treated plants and soils at 4 months p.i.(Table 4). This �RSL1 stability (Fig. 4) in soil appears to beanother advantage of �RSL1 as a biocontrol agent. Moreover,bulk production at high concentrations of �RSL1 particles(�1012 PFU/ml) is possible by centrifugation at 15,800 � g for20 min. Prolonged disease control may be possible if �RSL1 isapplied to plants at the seedling stage.

ACKNOWLEDGMENTS

This study was supported in part by the Industrial Technology Re-search Grant Program (04A09505) from the New Energy and Indus-trial Technology Development Organization of Japan and by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science, andTechnology of Japan (21580095 to T.Y. and 20.3533 to A.F.).

REFERENCES

1. Abedon, S. T. 2009. Disambiguating bacteriophage pseudolysogeny: a histor-ical analysis of lysogeny, pseudolysogeny, and the phage carrier state, p.285–307. In H. T. Adams (ed.), Contemporary trends in bacteriophage re-search. Nova Science Publishers, Inc., Hauppauge, NY.

2. Ausubel, F., et al. 1995. Short protocols in molecular biology, 3rd ed. JohnWiley & Sons, Inc., Hoboken, NJ.

3. Balogh, B., J. B. Jones, F. B. Iriarte, and M. T. Momol. 2010. Phage therapyfor plant disease control. Curr. Pharm. Biotechnol. 11:48–57.

4. Fegan, M., and P. Prior. 2005. How complex is the Ralstonia solanacearumspecies complex? p. 449–461. In C. Allen, P. Prior, and A. C. Hayward (ed.),Bacterial wilt: the disease and the Ralstonia solanacearum species complex.American Phytopathology Society, St. Paul, MN.

5. Fujie, M., H. Takamoto, T. Kawasaki, A. Fujiwara, and T. Yamada. 2010.Monitoring growth and movement of Ralstonia solanacearum cells harboringplasmid pRSS12 derived from bacteriophage �RSS1. J. Biosci. Bioeng. 109:153–158.

6. Fujiwara, A., T. Kawasaki, S. Usami, M. Fujie, and T. Yamada. 2008.Genomic characterization of Ralstonia solanacearum phage �RSA1 and itsrelated prophage (�RSX) in strain GMI1000. J. Bacteriol. 190:143–156.

7. Gill, J., and S. T. Abedon. 2003. Bacteriophage ecology and plants. AmericanPhytopathological Society, St. Paul, MN. http://apsnet.org/online/feature/phages/.

8. Goto, N. 1992. Fundamentals of bacterial plant physiology. Academic Press,Inc., New York, NY.

9. Hayward, A. C. 1991. Biology and epidemiology of bacterial wilt caused byPseudomonas solanacearum. Annu. Rev. Phytopathol. 29:65–87.

10. Hayward, A. C. 2000. Ralstonia solanacearum, p. 32–42. In J. Lederberg (ed.),Encyclopedia of microbiology, vol. 4. Academic Press, Inc., San Diego, CA.

11. Horita, M., and K. Tsuchiya. 2002. Causal agent of bacterial wilt diseaseRalstonia solanacearum, p. 5–8. In National Institute of Agricultural Sciences(ed.), MAFF microorganism genetic resources manual no. 12. National In-stitute of Agricultural Sciences, Tsukuba, Japan.

12. Ignoffo, C. M., and C. Garcia. 1992. Combination of environmental factorsand simulated sunlight affecting activity of inclusion bodies of the heliothis(Lepidoprera: Noctuidae) nucleopolyhedrosis virus. Environ. Entomol. 21:210–213.

13. Iriarte, F. B., et al. 2007. Factors affecting survival of bacteriophage ontomato leaf surfaces. Appl. Environ. Microbiol. 73:1704–1711.

14. Janse, J. 1996. Potato brown rot in western Europe-history, present occur-rence and some remarks on possible origin, epidemiology, and control strat-egies. Bull. OEPP/EPPO 26:679–695.

15. Jones, J. B., et al. 2007. Bacteriophages for plant disease control. Annu. Rev.Phytopathol. 45:245–262.

16. Kawasaki, T., H. Satsuma, M. Fujie, S. Usami, and T. Yamada. 2007.Monitoring of phytopathogenic Ralstonia solanacearum cells using greenfluorescent protein-expressing plasmid derived from bacteriophage �RSS1.J. Biosci. Bioeng. 104:451–456.

17. Kawasaki, T., et al. 2009. Genomic characterization of Ralstonia so-lanacearum phage �RSB1, a T7-like wide-host-range phage. J. Bacteriol.191:422–427.

18. Okabe, N., and M. Goto. 1963. Bacteriophages of plant pathogens. Annu.Rev. Phytopathol. 1:397–418.

19. Ozawa, H., H. Tanaka, Y. Ichinose, H. Shiraishi, and T. Yamada. 2001.Bacteriophage P4282, a parasite of Ralstonia solanacearum, encodes a bac-teriolytic protein important for lytic infection of its host. Mo. Genet. Genom.265:95–101.

20. Sambrook, J., and D. W. Russell. 2001. Molecular cloning: a laboratory manual,3rd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

21. Tanaka, H., H. Negishi, and H. Maeda. 1990. Control of tobacco bacterialwilt by an avirulent strain of Pseudomonas solanacearum M4S and its bac-teriophage. Ann. Phytopathol. Soc. Japan 56:243–246.

22. Toyoda, H., et al. 1991. Characterization of deoxyribonucleic acid of virulentbacteriophage and its infectivity to host bacterium, Pseudomonas so-lanacearum. J. Phytopathol. 131:11–21.

23. Yamada, T., et al. 2007. Isolation and characterization of bacteriophages thatinfect the phytopathogen Ralstonia solanacearum. Microbiology 153:2630–2639.

24. Yamada, T., et al. 2010. A jumbo phage infecting the phytopathogen Ral-stonia solanacearum defines a new lineage of the Myoviridae family. Virology398:135–147.

4162 FUJIWARA ET AL. APPL. ENVIRON. MICROBIOL.

on June 2, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from